On Tue, 19 Aug 2008 09:10:19 +0200, Øyvind Harboe wrote: >>> Can you try to boil down these discussions to the essential and >>> relevant bits for OpenOCD & post it here? >> >> I thought my previous post did exactly that. > > I guess there is no royal road to solving this problem... There is a lot > of necessary complexity to the situation that just need to be grokked.
Yes, and so far, nobody involved seems to have sufficient knowledge or ideas to solve the problem, with the possible exception of Nicolas Pitre. I am hoping he can weigh in. >> Here is some BASIC-inspired >> pseudocode that may alternately explain the situation: >> >> 1 call power_on >> 2 call sys_reset > > what is sys_reset? System reset of the target via the sRST line. >> 3 if flash_contents <> valid_executable_code then exception_asserted=-1 > > how would one check that flash does not contain valid executable code? The target CPU attempts to execute bootstrap code from flash upon power- on/reset, and it's not there. >> 4 if exception_asserted=0 then if halt_requested then goto 8 >> 5 if exception_asserted=0 then goto 3 >> 6 if you_feel=like_trying_again then goto 2 >> 7 print "I give up!":end >> 8 call write_to_the_flash >> >> Do you see the problem? > > No. I guess I would be loathe to try to figure this out without access > to hardware and time to read up on the documentation and a scope even > perhaps. The target only checks whether OpenOCD wants it to halt in line 4, but line 3 prevents that from happening, in the above illustration of what happens. _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
